Analysis of GALFACTS Data for the Study Of Variable Radio Sources Scott Barenfeld 1, Tapasi Ghosh 2, Chris Salter 2 1 NAIC/University of Rochester, 2 NAIC Abstract The G-ALFA Continuum Transit Survey (GALFACTS) is a spectro-polarimetric survey of Arecibo Observatory’s visible sky from MHz, using the Arecibo L-band Feed Array (ALFA). Among the survey’s many scientific goals is a large-scale statistical study on the short-term variability of the flux density and polarization of radio sources. Every point in the sky is observed twice, with less than a month between observations, making this the largest systematic search for variability ever conducted. In this poster, we present the development of computer code to aid in this search and some preliminary results from this code. The code takes GALFACTS data in the form of time series for 2048 individual spectral channels containing positions and full-Stokes antenna temperatures, and turns these into a list of individual radio sources with their positions and Stokes-I temperatures. We first ran the code for the field surrounding the radio source S , of known flux density, as a test. Once a working code was completed, it was run on the field of another radio source, S References and Acknowledgements Bogdan Georgescu, Ilan Shimshoni, Peter Meer, "Mean Shift Based Clustering in High Dimensions: A Texture Classification Example," iccv, vol. 1, pp.456, Ninth IEEE International Conference on Computer Vision (ICCV'03) - Volume 1, 2003 Taylor, Andrew R. and Christopher J. Salter. “The G-ALGA Continuum Transit Survey.” arXiv: (2005) The Gaussian fitting code used was provided by J. Chengalur, GMRT, NCRA, India.. Remove NaNsOriginal Data Derive Average Spectrum Convolve with HPF in Frequency Find and Remove RFI Average Over Good Channels Time Series Noise Filter Time Series Convolution with Matched Source Fn Identify Individual Source Cuts Fit Gaussians to Source Deflection Generate File of Source Cuts Repeat for Each Observing Day Group Cuts from Different Days Estimate Source Flux Densities Sort Source File Average Spectrum- This plot shows the average Stokes-I values for each frequency channel after NaNs have been removed. HPF Convolution- The averages by channel have been brought to a baseline of zero using convolution, making it easier to find RFI. Time Series- The channels with no RFI are used to generate an average Stokes-I value for each time data is recorded (every 200 ms). Gaussian Fit- This is the result of our fit of the large source in the middle of the time series shown left. Repeat for Each Day- This plot shows the distribution of cuts from each day of data. Group Cuts- Using mean-shift clustering, cuts from different days were grouped into individual sources. S Results This Gaussian fit was done to a section of the cuts of S The peak antenna temperature of 11.1 K and gain of Arecibo’s telescope (10 K/Jy) give an estimated flux density of 1.1 Jy. The authors (from left to right), Scott Barenfeld, Tapasi Ghosh, and Chris Salter.